Jan. 3, 1878] 



NATURE 



181 



reasoning as that by which the induction of currents is deduced 

 from the force exerted between a circuit and a magnet and the 

 existence of contact electromotive force from the Peltier effect, 

 it follows that a current should exist if two zinc electrodes con- 

 nected by a wire are immersed in a solution of sulphate of zinc, 

 the direction of the current being (in the solution) from the upper 

 to the lower electrode. 



" I tested this a few days ago, using a glass tube eighteen 

 inches long, filled with a saturated solution of 

 sulphate of copper and closed by copper caps 

 with wires attached. 



"On connecting the wires with a very 

 delicate Thomson's astatic galvanometer be- 

 longing to Prof. Halford, a very considerable 

 deflection was produced (200 divisions) when 

 the tube was held veriically, the direction of 

 the deflection being reversed when the tube was reversed. 



" If the tube, alter being held vertically, was placed in a hori- 

 zontal position, the deflection diminished, but several minutes 

 elapsed before the index came to zero, which it eventually did. 

 I cannot explain the time taken. I am now preparing to test the 

 actual loss of weight of the upper electrode. 

 •' I have the honour to be, Sir, 



** Your obedient servant, 



"F. J. Pi RANI, 



" Lecturer on Natural Philosophy and Logic, 

 University of Melbourne. 



" P.S. — If the phenomenon has not been noticed before I shall 

 be obliged if you will kindly communicate it to Nature, 



" F. J. P." 



The Telephone 



I HAVE been much interested in the communication by Dr. 

 Rontgen on a telephonic alarum. During the past six or seven 

 weeks, in investigating the phenomena of the telephone, chiefly as 

 to the suggestions they offer regarding the mechanism of nervous 

 transmission, I have frequently shown to friends the striking 

 experiment described by Dr. Rontgen, and, amongst others, 

 to Sir William Thomson. It has succeeded with Ut^, 

 Ut^, and with numerous forks up to Ut^, but, as stated by 

 Dr. Rontgen, the best result was obtained with Ut,^. With 

 those below this pitch the tone was feeble, whilst with those 

 above it it was transient, in consequence of the difficulty of 

 keeping the small fork going. With Ut^, worked continuously 

 by an electro-magnet, another fork of the same pitch sounded 

 loudly and steadily. I have also been engaged in some endea- 

 vours to record on a moving surface the vibrations of the plate. 

 These have been so successful as to show that it is only a question 

 of delicate adjustment. In endeavouring to utilise one telephone 

 by making several friends listen at once, I have found that by 

 fixing the metal disc to a thin membrane over a small cavity 

 filled with air, like a Koenig's capsule, and having a number of 

 flexible leaden tubes connected with it, an ear placed at the end 

 of each tube will hear distinctly. John G. McKendrick 



Physiological Laboratory, University of Glasgow, 

 December 31, 1877 



The Radiometer and its Lessons 



Prof. Osborne Reynolds (vol. xviii. p. 121) appears to 

 have done himself less than justice in the extracts he has sent 

 you from his earlier papers, as representing his published views 

 on the action of residual gas in radiometers. For the extracts do 

 not suffice to constitute an explanation of this action, whereas 

 the papers from which he makes the extracts contained what, 

 if true, might have been an explanation cf the action of residual 

 gas, along with much else that is admittedly erroneous ; and 

 although those papers (the only ones published before mine) con- 

 clude with Prof. Reynolds's own expression of opinion that 

 residual gas is not the cause of the force observed by Mr. Crookes. 



He quotes three paragraphs. In two of these he recited the 

 fundamental principle in the kinetic theory of gases which he 

 sought to apply. To obtain an explanation of the phenomenon 

 from this principle according to the method pursued by Prof. 

 Reynolds, it was necessary for him {a) to establish a law con- 

 necting an excess of force perpendicular to the disc with 

 a flow of heat in radiometers, and {b) to indicate agencies 

 which could occasion a sufficient flow of heat. He quotes 

 the passage in which he announced the result of his, as I 



believe, unsuccessful attempt to accomplish the former of 

 these, but he. omits the equally necessary passage in which he 

 dealt with the latter. It will be found at page 407 of the 

 Proceedings of the Royal Society, vol. xxii., and is couched in 

 the following terms : — " It must be remembered that e [which 

 measures the outflow of heat] depends on the rate at which 

 cold particles will come up to the hot surface, which is very slow 

 when it depends only on the diffusion of the particles of the gas 

 inter se, and the diffusion of the heat among them. It will be 

 much increased by convection currents." If this passage, as was 

 requisite, had been added to the extracts made by Prof. Reynolds, 

 it would have brought his recent account of the views he had 

 announced into conformity with my account of them. 



In connection with this subject it should be observed that 

 Prof. Osborne Reynolds has in express terms excluded from his 

 explanation that which I believe to be the real agency which 

 brings a sufficient supply of cold molecules up to the hot surface, 

 for he states, in his letter to Nature (vol. xvii., p. 27), that " it 

 is incompatible with his explanation that the increase resulting 

 from rarefaction in the mean length of the path of the gaseous 

 molecules would favour the action." Now the polarisation of 

 the gas depends on the ratio which this mean length bears to the 

 interval between heater and cooler. 



I cannot find anywhere in Prof. Osborne Reynolds's writings 

 an explanation of the thing to be explained, viz., that the stress 

 in a Crookes's layer is different in one direction from what it is 

 at right angles to that direction. Let v be the component of the 

 momenta of the molecules striking a square unit of the heater in 

 the unit of time, resolved perpendicularly towards the heater ; 

 and let u be the corresponding normal component of their 

 momenta from the heater, when they are thrown off. Then 

 u + V IS the pressure on the heater. Now if u and v could 

 result respectively from unpolarised motions in the gas, the 

 momentum resolved parallel to the heater would be \u -^ \v 

 from left to right, with an equal momentum from right to left. 

 Adding these we find « + z/ the pressure of the gas parallel to 

 the heater. This is equal to the normal pressure, and, therefore, 

 under these circumstances, there would be no Crookes's force 

 whatever. It is only when we take the polarisation of the gas 

 into account that the momenta resolved, parallel to the heater 

 become different from \u and \v. 



Prof. Osborne Reynolds says that my views are at variance 

 with results arrived at by Clausius and other discoverers in 

 this branch of physic?. I do not myself value appeals to 

 authority in matters of science. But it so happens that here 

 again it appears to be Prof Reynolds who makes the mistake. 

 Clausius, in his great memoir on the conduction of heat by 

 gases, published in 1862 {Phil. Mag., vol. xxiii. p. 529), warns 

 his readers against the very error into which Prof. Reynolds seems 

 to fall, and points out that there "are obvious limits" be)ond 

 which the laws he had discovered for the conduction of heat do 

 not prevail, one of which limits is that the gas ' ' must not be so 

 expanded that the mean length of excursion of the molecules 

 becomes so great that its higher powers cannot be neglected." 

 Now it is just to this excepted case, to the Sprengel vacua ex- 

 perimented on by Mr. Crookes, that Prof. Osborne Reynolds 

 applies the laws of conduction, and he then objects to my theory 

 that it does not agree witfi the laws so misapplied. The 

 phenomenon of Crookes's stress appears to come into exist- 

 ence precisely in Clausius's excepted case, viz., so soon as the 

 ratio which the mean length of excursion of the molecules 

 bears to the interval between heater and cooler, is such, that 

 when multiplied by a function of the temperatures of the heater 

 and cooler, its square is of appreciable magnitude in Clausius's 

 equations. This may be experimentally secured either by placing 

 the heater and cooler very close together, as in experiments upoa 

 spheroidal drops, or by excessively attenuating the gas so as to 

 lengthen the free paths of the molecules sufficiently, as in 

 radiometers. G. Johnstone Stoney 



Dublin, December 20 



Postscript, December 22.— I have just seen Prof Schuster's 

 letter (Nature, vol. xvii. p. 143). Dr. Schuster will pardon 

 me if I say that he has adopted a scarcely legitimate course in 

 introducing into a discussion on priority his present reminiscence 

 of one of the conversations about the radiometer which he held 

 with his friend, Prof. Osborne Reynolds, two and a half years 

 ago. The language in which he reports it is foreign to Prof. 

 Reynolds's style of composition, so that we may conclude we 

 are dealing with Dr. Scfiuster's words, and the words which 

 occurred to him after he had read much else en the subject. No 



